Notes for AA214, Chapter 2

T. H. Pulliam
Stanford University

1
 Wave Equation

Continuous PDE: xmin ≤ x ≤ xmax, with a = constant

∂ u(x, t) ∂ u(x, t)
+a =0 (1)
∂t ∂x
1. PDE Theory requires an Initial Condition (IC) and Boundary
Conditions (BC)
2. IC: u(x, 0) = g(x), an arbitrary function of x, must satisfy BC
3. BC: The first order PDE in x requires only one BC, satisfying IC
(a) If a ≥ 0, then u(xmin, t) = l(t)
(b) If a < 0, then u(xmax, t) = r(t)

2
 Discussion of BC: Non-Periodic

1. Scalar quantity u is given on one boundary, corresponding to a

wave entering the domain thru this “inflow” boundary.
(a) No boundary condition is specified at the opposite side, the
“outflow” boundary.
(b) This is consistent in terms of the well-posed-ness of a
first-order PDE.
(c) Hence the wave leaves the domain through the outflow
boundary without distortion or reflection.
(d) Note that the left-hand boundary is the inflow boundary
when a is positive, while the right-hand boundary is the
inflow boundary when a is negative.

3
 Discussion of BC: Periodic

1. The flow being simulated is periodic.

(a) At any given time, what enters on one side of the domain
must be the same as that which is leaving on the other.
(b) This is referred to as the biconvection problem.
(c) It is the simplest to study and serves to illustrate many of the
basic properties of numerical methods applied to problems
involving convection, without special consideration of
boundaries.
(d) We pay a great deal of attention to it in the initial lectures.

4
 Periodic Wave Equation

1. Next we study the properties of the Periodic Wave Equation

∂ u(x, t) ∂ u(x, t)
+a =0, 0 ≤ x ≤ 2π (2)
∂t ∂x
2. BC: u(0, t) = u(2π, t)
3. IC: u(x, 0) = g(x), g(0) = g(2π)

5
 Periodic Wave Form

1. The general solution to Eq.1 is:

u(x, t) = g(x − at)

with g(x) satisfying the IC

2. We will choose a specific form of the solution for periodic flow
3. Fourier Series: An Arbitrary Periodic (Harmonic) Function Can
Be Represented By A Fourier Series
M
X X
iκm x
g(x) = fm (0)e = gm (x) (3)
m=−N m

6
 Examples of Periodic Fourier Functions

1. Simple Sine
eix − e−ix
sin(x) =
2i
M, N = 1, κ1 = 1, κ−1 = −1
1 −1
f1 (0) = , f−1 (0) =
2i 2i
2. Sum of Sine and Cosine
e3ix − e−3ix e5ix + e−5ix
2.0sin(3x) + 0.1cos(5x) = 2.0 + 0.1
2i 2
M, N = 5, κ3 = 3, κ−3 = −3, κ5 = 5, κ−5 = −5,
2.0 −2.0 0.1 0.1
f3 (0) = , f−3 = , f5 (0) = , f−5 =
2i 2i 2 2

7
 Linear Superposition Theory

1. Equation 1 is a linear equation in u(x, t) and must satisfy an

arbitrary g(x) from Eq.3
2. By the Theory of Linear Superposition, given two or more
solutions, e.g., u1 (x, t), u2 (x, t)
(a) If u1 (x, t) Satisfies Eq.1 and u2 (x, t) Satisfies Eq.1
(b) Then: The sum of u(x, t) = u1 (x, t) + u2 (x, t) also satisfies
Eq.1

8
 Generalize Solution

1. Eq.3 is a sum of various periodic functions eiκm x , each of which

taken separately leads to general solutions um (x, t) = gm (x − at)
(a) Simplify and generalize our solutions class by choosing the
general g(x) = eiκx
(b) Consider each wave component separately, (ie. general κ)
2. General Solution for Periodic IC
M
X
u(x, t) = fm (0)eiκm (x−at) (4)
m=−N

9
Separation of Variable Solution of Wave Equation

1. Using separation of variables assuming a general form

u(x, t) = eiκx f (t)

(arbitrary κ)
∂ u(x,t)
2. Apply the general result ∂x = iκ u(x, t) to Eq.2,

∂ u(x, t) ∂ u(x, t)
+a =0
∂t ∂x
∂ eiκx f (t)
+ aiκeiκx f (t) = 0
∂t

10
 PDE - ODE

1. The ODE for f (t) is

∂ f (t)
+ a iκ f (t) = 0
∂t
with solution
f (t) = f (0)e−aik t
giving
u(x, t) = ceiκx e−aiκ t , c = f (0)

2. So the General Solution to Eq.2, (for each κ),

u(x, t) = ceiκ(x−a t) (5)

11
 General Solution

X
u(x, t) = cm eiκ(x−a t) (6)
m

1. This will be the exact solution which we will use to evaluate the
effects of
∂u
(a) Approximating ∂x with Numerical Finite Differences.
∂u
(b) Approximating ∂t with Various Time Advance Schemes.

12